Genetic divergence, using Mahalanobis D2 statistic, was worked out in 30 traditional upland rice varieties from nine states of India including the Northeastern Region. Based on 10 agromorphological characters, these varieties were grouped into six clusters. Cluster I was the largest with 66.6% genotypes from different states except Punjab. Clusters IV, V and VI were monogenotypic. Varieties from the Northeastern Region exhibited the maximum diversity, and were represented in all the clusters except cluster V. However, no parallelism was observed between geographic diversity and genetic diversity. Characters like No. of secondary branches/panicle, yield/plant, and No. of fertile grains/panicle had sizeable contribution to total divergence. These characters could, therefore, form the basis for selection of parents from distantly placed clusters to obtain high heterotic combinations.
Upland rice, genetic diversity
